Insufficiency of Some Experiments 189 



one means of gaming knowledge, these institutions, 

 and to some extent private farmers, have conducted 

 many so-called practical feeding experiments in order 

 to verify present beliefs, test theories and solve exist- 

 ing problems. The relative value of various feeding 

 stuffs and rations for producing growth and milk and 

 the influence of different fodders and grain foods upon 

 the quality of the product have been the subjects of 

 numerous feeding tests. Much valuable information 

 has been secured in this way, but there has not 

 always been a full recognition, even by experiment 

 stations, of the limitations which should be observed 

 in drawing conclusions from this manner of experi- 

 mentation. 



In order to view this matter more in detail, let us 

 consider experiments in testing rations for growth and 

 milk production. The usual method of procedure with 

 such feeding trials is either to feed two lots of animals 

 on the rations to be compared and note the comparative 

 growth or milk yield, or to feed the same lot on one 

 ration for a time and then change to another ration. 



If these tests are made with growing or fattening 

 animals, the increase in live weight is taken as the 

 measure of the relative efficiency of the rations com- 

 pared. It should be said of these experiments that 

 their apparent verdict is to be accepted with great 

 caution, and definite conclusions are not justified until 

 repeated trials of two rations or of two systems of 

 feeding, made with the use of all possible precautions 

 against error, and under a variety of conditions, give 

 uniform and consistent results in the same direction. 



